The PyTorch implementation for the paper titled "An Efficient Data-Driven Approximation to the Stochastic Differential Equations with Non-global Lipschitz Coefficient and Multiplicative Noise." by X. Qi, T. Duan, and H. Guo.

The contribution this paper is to propose a neural approximation called "extended continuous latent process flow" for numerically solving underlying model. The principle idea of this method is to derive a variational lower bound by constructing a posterior latent process conditional on all information over the whole time interval to maximize the log-likelihood generated by the observations, thereby providing a feasible way to approximate the considered problem. Numerical experiments are reported to demonstrate the effectiveness and generalization performance of the proposed method.

Create new environment for this code

conda create -n ECLPF
conda activate ECLPF
pip install -r requirements.txt

python eCLPF1.py

python eCLPF1.py